Websites that implement community features provide users the ability to present content in association therewith for consumption by all users. Examples of this user-created content are product reviews and discussion group posts. Websites typically contain a web page with guidelines that users are asked to adhere to when submitting content for presentation. Some users, either intentionally or unintentionally, violate published guidelines by submitting content for presentation that is abusive due to profanity, spam, or inclusion of other content that violates the website's guidelines for user-created content. Presentation of this abusive content on the website can lead to customer dissatisfaction, distrust, or even legal issues if the content is egregious enough (e.g., if the user-created content includes threats, racist language, or the like).
Existing processes for detecting and preventing publication of abusive user-created content are either manual processes that involve a large number of support staff reading content in its entirety, or automatic processes with overly simplistic logic that users can easily circumvent by slightly obfuscating potentially abusive language.